Dr. Marko Pejković

After the introduction, the first part of the paper is devoted to defining the concepts of artificial intelligence and totalitarianism, where the importance of distinguishing between the current (machine learning) and the projected (superintelligence) phase in the development of artificial intelligence, i.e. between the embryonic (totalitarian movement out of power) and the established (totalitarian movement in power) stage in the development of totalitarianism is underlined. The second part of the paper examines the connection between the current level of artificial intelligence and the embryonic phase of totalitarianism, while the third part of the paper analyzes the potential relationship between the superintelligence and the established totalitarianism. It seems, considering the similarities and differences between the effects of contemporary and future artificial intelligence and the effects of earlier totalitarianism, that today (and in the future) we do not have a mere replica of totalitarian phases from the 20th century, but special totalitarian phenomena in the form of "capillary totalitarianism", i.e. "hypertotalitarianism". Last century's totalitarianism, as well as today's "capillary" variant of it, were not necessarily irreversible, but "hypertotalitarianism" will be. In conclusion, protective measures against the risk of artificial intelligence are proposed, in the form of the principle of exemption (modeled after the concept of conscientious objection).

The aim of this paper is to mesure the degree of executive stability in Serbia via possible correlation between two political notions: one being the fragmentation of the party system and the other the stability of the government, e.g. the length of the period during which the “cabinet” is in office. If the correlation is strong, we can conclude that this would not only permit us to follow further changes of their relationships, so to have some basic insights of their processual tendencies also, but it would also enable us to raise the Serbian political science to a higher level, which implies a relatively reliable prognostics through the use of the regression. Prior to that, an overview of all known formulas for the fractionalization is presented and one option is chosen, for even though the formula does not influence the coefficient of the correlation or the regression, we have wanted to clarify some counterintuitive inconsistencies which can emerge when bias toward under/overestimation of the fragmentation in the formulas is mathematically detected.